CN107818199A - Mountain area 10kV line-to-ground resistive arrangement methods based on adaptive neural network - Google Patents
Mountain area 10kV line-to-ground resistive arrangement methods based on adaptive neural network Download PDFInfo
- Publication number
- CN107818199A CN107818199A CN201710888308.8A CN201710888308A CN107818199A CN 107818199 A CN107818199 A CN 107818199A CN 201710888308 A CN201710888308 A CN 201710888308A CN 107818199 A CN107818199 A CN 107818199A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- sampling instant
- formula
- munderover
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
Abstract
The present invention discloses a kind of mountain area 10kV line-to-ground resistive arrangement methods based on adaptive neural network.Measured using soil resistivity table and obtain the stratified soil thickness in region residing for 10kV overhead line structures and the multiple sample datas of layered resistance rate, input signal using sample data as adaptive neural network, using equivalent soil resistivity as output signal, connection weight between adaptive neural network input layer, hidden layer and output layer is optimized by swarm optimization method, equivalent soil resistivity optimal value is obtained after iterative learning, then the grounding resistance after being optimized by grounding resistance design formula.Using the accurate calculating of 10kV overhead transmission line shaft tower grounding resistances under the achievable mountain area complicated geological environment of the present invention, accurately design and manufacture bases are provided for 10kV overhead transmission line lightening protection engineerings.With the advantages below not available for prior art:Computational accuracy is higher, and computational methods Project Realization is simpler.
Description
Technical field
The present invention relates to the grounding technology of field of power, more particularly to a kind of mountain area based on adaptive neural network
10kV line-to-ground resistive arrangement methods.
Background technology
Grounding Technology of Modern Power System is to ensure power system and its pass of electrical equipment safe and reliable operation and personal security
One of key technology.At present, the instrument of most measurement soil resistivities can not measure the soil electricity of different depth level simultaneously
Resistance rate, can not monitor the soil resistivity value of multiple spot incessantly for a long time, thus project planner be difficult to it is straight by instrument
The mode for connecing measurement obtains the real soil resistivity value of complicated geological environment.At present, industry is mainly based upon power system and connect
The related state's network planning journey in ground, numerical computations, main flow are carried out to the ground connection parameter of earthed system with reference to Computer-aided Design Technology
Method include FInite Element, compound image method, CDEGS (Current Distribution, Electromagnetic
Fields, Grounding and Soil Structure Analysis) software analysis method, but because current field is distributed in nothing
In the big complicated soil space of limit, have for the grounding resistance optimization design under complicated geological environment extremely challenging.
Neutral net has the advantages that stronger non-linear mapping capability, generalization ability and preferable fault-tolerance, therefore
It is widely used, but is being grounded in fields such as pattern-recognition, System Discrimination, image procossing, optimization calculating, optimum controls
Also rarely has application in resistance optimization design.On the other hand, the parameter to be optimized adjusted is more in neutral net, a large amount of connection weights
The design of parameter relies primarily on the design experiences and test of many times of designer at present, thus how to design one towards specific work
Journey item purpose adaptive neural network is as one of domestic and international academia and engineer applied field problem urgently to be resolved hurrily.
The content of the invention
In view of the above-mentioned deficiencies in the prior art, it is an object of the present invention to provide the mountain area 10kV lines based on adaptive neural network
Road grounding resistance design method.
In order to solve the above-mentioned technical problem, the present invention is achieved by the following technical solutions:Based on adaptive neural network net
The mountain area 10kV line-to-ground resistive arrangement methods of network, this method comprise the following steps:
A, for 10kV overhead transmission lines to be designed residing for geological environment, using soil resistivity table measure obtain the region
The sample data of diverse location, including upper layer of soil thickness tu, upper layer of soil electricalresistivityρu, BC soil thickness tm, BC soil
ElectricalresistivityρmWith lower soil electricalresistivityρl, wherein N is the maximal dimension of sample;
Adaptive neural network parameter values, including input layer number n are set B,I, output layer nodes nO, hidden layer
NodesRepresent (nI+nO)0.5Round up number, population scale M, maximum
Iterations Imax, crossover probability pc, mutation probability pm;
C, the initialization population P of real coding is generated at random0={ Sh, h=1,2 ..., M }, wherein Sh=(wij(1),wjl
(1)), i=1,2 ..., nI, j=1,2 ..., nH, l=1,2 ..., nO, wij(1) the 1st sampling instant input node i is represented
With the connection weight between hidden layer node j, wjl(1) represent the 1st sampling instant hidden layer node j and output node layer l it
Between connection weight;
D, P in assessing0The fitness of each individual;
E, by { F (Sh), h=1,2 .., M } sorted according to ascending power, by the minimum individual mark of fitness value for it is current most
Good individual Sbest, by the fitness value of minimum labeled as current preferably target function value Fbest;
F, according to the cumulative probability shown in formula (8) from P0Selection produces new colony P;
Wherein, p (Sh) represent individual ShCumulative probability;
G, according to crossover probability pcTwo individual S are randomly choosed out from PiAnd Si+1, intersect according to formula (9) and produce new two
Individual SniAnd Sn(i+1), SniAnd Sn(i+1)Original S is replaced respectivelyiAnd Si+1, other individuals keep constant in P, new by now
Population is labeled as Pc={ Sch, h=1,2 ..., M }, wherein SchRepresent PcIn h-th individual;
Caused random number between wherein α represents 0 to 1;
H, according to mutation probability pmFrom PcThe middle individual S of selectionch, according to formula (10) to SciRandom variation is carried out, is produced new
Individual SNh, and by SNhInstead of Sch, keep PcIn other individuals it is constant, so as to produce new population PN;
SNh=Sch+(r-0.5)×max{2(Sch-L),2(U-Sch)} (10)
Wherein L and U represent S respectivelychLower and upper limit, r represent 0 to 1 between caused random number;
I, P is unconditionally received0=PN;
J, repeat step D to step I is until meeting largest optimization iterations Imax;
K, the current preferably target function value F of outputbestWith best individual Sbest, and equivalent soil resistivity corresponding to output
Predicted value ρ=yOl(Tmax);
L, 10kV overhead transmission line shaft tower grounding resistances R is calculated according to formula (11);
Wherein, ρ is the soil resistivity obtained in step (11), and L is grounded for 10kV overhead transmission lines Reinforced Concrete Pole Tower
The total length of pole, h be earthing pole the depth of burying, d be earthing pole diameter, AtTo be grounded the form factor of level.
Preferably, the step D assesses P in accordance with the following steps0In each individual fitness:
D1, make k=1;
D2, by tu、ρu、tm、ρmAnd ρlAs the input signal of adaptive neural network, by the input of kth time sampling instant
Signal XI(k) it is defined as XI(k)={ xi(k), i=1,2 ..., nI, wherein nIFor neural network input layer nodes, in this nI
=5, x1(k)=tu,x2(k)=ρu,x3(k)=tm,x4(k)=ρm,x5(k)=ρl;
D3, the neutral net hidden layer input signal X for calculating kth time sampling instantHI(k)={ xHIj(k), j=1,
2,...,nH, wherein xHIj(k) calculate as shown in formula (1):
Wherein, wij(k) connection weight of the kth time between sampling instant input node i and hidden layer node j is represented;
D4, the neutral net hidden layer output signal X for calculating kth time sampling instantHO(k)={ xHOj(k), j=1,
2,...,nH, wherein xHjO(k) calculate as shown in formula (2):
D5, the neutral net output layer signal Y for calculating kth time sampling instantO(k)={ yOl(k), l=1,2 ..., nO,
Wherein nORepresent output layer number of nodes, yOl(k) calculate as shown in formula (3):
Wherein, wjl(k) connection weight between kth time sampling instant hidden layer node j and output node layer l is represented;
D6, assess according to formula (4) kth time sampling instant neutral net error performance index J (k);
Wherein, yOlr(k) desired value of the equivalent soil resistivity of kth time sampling instant is represented;
D7, the weight coefficient w according to formula (5) renewal+1 sampling instant of kthij(k+1);
Wherein, η represents learning rate, and β represents factor of momentum;
D8, the weight coefficient w according to formula (6) renewal+1 sampling instant of kthjl(k+1), and k=k+1 is made;
wjl(k+1)=wjl(k)+η(yOlr(k)-yOl(k))xHOj(k)+β(wij(k)-wij(k-1)) (6)
D9, repeat step D2 are to step D8 until meeting k=Tmax, wherein TmaxRepresent maximum sampling number;
D10, according to formula (7) calculate individual ShFitness F (Sh);
Compared with prior art, it is an advantage of the invention that:Using the achievable mountain area complicated geological environment 10kV framves of the present invention
The accurate calculating of ceases to be busy line pole tower grounding resistance, accurately design and manufacture bases are provided for 10kV overhead transmission line lightening protection engineerings.
With the advantages below not available for prior art:Computational accuracy is higher, and computational methods Project Realization is simpler.
Brief description of the drawings
Fig. 1 is the mountain area 10kV line-to-ground resistive arrangement Method And Principle schematic diagrames based on adaptive neural network.
Embodiment
Embodiments of the invention are described below in detail, the example of the embodiment is shown in the drawings.Below with reference to
The embodiment of accompanying drawing description is exemplary, it is intended to for explaining the present invention, and is not considered as limiting the invention.
Refering to the implementation that Fig. 1 is the mountain area 10kV line-to-ground resistive arrangement methods of the invention based on adaptive neural network
Example, by taking Zhejiang Nanshan District 10kV overhead transmission lines as an example, using the mountain area 10kV proposed by the present invention based on adaptive neural network
Line-to-ground resistive arrangement method is implemented.
A, for 10kV overhead transmission lines to be designed residing for geological environment, obtained using ES3010E soil resistivity table measurements
Obtain the sample data of the region diverse location, including upper layer of soil thickness tu, upper layer of soil electricalresistivityρu, BC soil thickness tm、
BC soil electricalresistivityρmWith lower soil electricalresistivityρl, wherein N is the maximal dimension of sample;
Adaptive neural network parameter values, including input layer number n are set B,I=5, output layer nodes nO=1,
Node in hidden layer, population scale M=30, maximum iteration Imax=50, crossover probability pc
=0.8, mutation probability pm=0.1;
C, the initialization population P of real coding is generated at random0={ Sh, h=1,2 ..., M }, wherein Sh=(wij(1),wjl
(1)), i=1,2 ..., 5, j=1,2 ..., 8, l=1, wij(1) the 1st sampling instant input node i and hidden layer section are represented
Connection weight between point j, wjl(1) represent the 1st sampling instant hidden layer node j and export the connection weight between node layer l
Weight;
D, P is assessed in accordance with the following steps0In each individual fitness;
D1, make k=1;
D2, by tu、ρu、tm、ρmAnd ρlAs the input signal of adaptive neural network, by the input of kth time sampling instant
Signal XI(k) it is defined as XI(k)={ xi(k), i=1,2 ..., nI, wherein nIFor neural network input layer nodes, in this nI
=5, x1(k)=tu,x2(k)=ρu,x3(k)=tm,x4(k)=ρm,x5(k)=ρl;
D3, the neutral net hidden layer input signal X for calculating kth time sampling instantHI(k)={ xHIj(k), j=1,
2,...,nH, wherein xHIj(k) calculate as shown in formula (1):
Wherein, wij(k) connection weight of the kth time between sampling instant input node i and hidden layer node j is represented;
D4, the neutral net hidden layer output signal X for calculating kth time sampling instantHO(k)={ xHOj(k), j=1,
2,...,nH, wherein xHjO(k) calculate as shown in formula (2):
D5, the neutral net output layer signal Y for calculating kth time sampling instantO(k)={ yOl(k), l=1,2 ..., nO,
Wherein nORepresent output layer number of nodes, yOl(k) calculate as shown in formula (3):
Wherein, wjl(k) connection weight between kth time sampling instant hidden layer node j and output node layer l is represented;
D6, assess according to formula (4) kth time sampling instant neutral net error performance index J (k);
Wherein, yOlr(k) desired value of the equivalent soil resistivity of kth time sampling instant is represented;
D7, the weight coefficient w according to formula (5) renewal+1 sampling instant of kthij(k+1);
Wherein, η represents learning rate, and β represents factor of momentum;
D8, the weight coefficient w according to formula (6) renewal+1 sampling instant of kthjl(k+1), and k=k+1 is made;
wjl(k+1)=wjl(k)+η(yOlr(k)-yOl(k))xHOj(k)+β(wij(k)-wij(k-1)) (6)
D9, repeat step D2 are to step D8 until meeting k=Tmax, wherein TmaxRepresent maximum sampling number;
D10, according to formula (7) calculate individual ShFitness F (Sh);
E, by { F (Sh), h=1,2 .., M } sorted according to ascending power, by the minimum individual mark of fitness value for it is current most
Good individual Sbest, by the fitness value of minimum labeled as current preferably target function value Fbest;
F, according to the cumulative probability shown in formula (8) from P0Selection produces new colony P;
Wherein, p (Sh) represent individual ShCumulative probability;
G, according to crossover probability pcTwo individual S are randomly choosed out from PiAnd Si+1, intersect according to formula (9) and produce new two
Individual SniAnd Sn(i+1), SniAnd Sn(i+1)Original S is replaced respectivelyiAnd Si+1, other individuals keep constant in P, new by now
Population is labeled as Pc={ Sch, h=1,2 ..., M }, wherein SchRepresent PcIn h-th individual;
Caused random number between wherein α represents 0 to 1;
H, according to mutation probability pmFrom PcThe middle individual S of selectionch, according to formula (10) to SciRandom variation is carried out, is produced new
Individual SNh, and by SNhInstead of Sch, keep PcIn other individuals it is constant, so as to produce new population PN;
SNh=Sch+(r-0.5)×max{2(Sch-L),2(U-Sch)} (10)
Wherein L and U represent S respectivelychLower and upper limit, r represent 0 to 1 between caused random number;
I, P is unconditionally received0=PN;
J, repeat step (4) to step (9) until meeting largest optimization iterations Imax=50;
K, the current preferably target function value F of outputbestWith best individual Sbest, and equivalent soil resistivity corresponding to output
Predicted value ρ=yOl(Tmax);
L, 10kV overhead transmission line shaft tower grounding resistances R is calculated according to formula (11);
Wherein, ρ is the soil resistivity obtained in step (11), and L is grounded for 10kV overhead transmission lines Reinforced Concrete Pole Tower
The total length of pole, h be earthing pole the depth of burying, d be earthing pole diameter, AtTo be grounded the form factor of level, in of the invention
Used Radiation coefficient is At=2.
By to the experimental result pair using the prior art such as the technology of the present invention and FInite Element, compound image method, CDEGS
Than analysis, we can be found that:Calculated using the mountain area complicated geological environment 10kV overhead transmission line shaft towers earth resistance meter of the present invention
Precision compared with prior art at least improve more than 2%, for 10kV overhead transmission line lightening protection engineerings provides accurately design and construction according to
According to.With the advantages below not available for prior art:Computational accuracy is higher, and computational methods Project Realization is simpler.
The specific embodiment of the present invention is the foregoing is only, but the technical characteristic of the present invention is not limited thereto, Ren Heben
The technical staff in field in the field of the invention, all cover among the scope of the claims of the present invention by the change or modification made.
Claims (2)
1. the mountain area 10kV line-to-ground resistive arrangement methods based on adaptive neural network, it is characterised in that:This method includes
Following steps:
A, for 10kV overhead transmission lines to be designed residing for geological environment, measure that to obtain the region different using soil resistivity table
The sample data of position, including upper layer of soil thickness tu, upper layer of soil electricalresistivityρu, BC soil thickness tm, BC soil resistance
Rate ρmWith lower soil electricalresistivityρl, wherein N is the maximal dimension of sample;
Adaptive neural network parameter values, including input layer number n are set B,I, output layer nodes nO, hidden layer node
Number Represent (nI+nO)0.5Round up number, population scale M, greatest iteration time
Number Imax, crossover probability pc, mutation probability pm;
C, the initialization population P of real coding is generated at random0={ Sh, h=1,2 ..., M }, wherein Sh=(wij(1),wjl(1)),
I=1,2 ..., nI, j=1,2 ..., nH, l=1,2 ..., nO, wij(1) represent the 1st sampling instant input node i with it is hidden
Connection weight between j containing node layer, wjl(1) represent between the 1st sampling instant hidden layer node j and output node layer l
Connection weight;
D, P in assessing0The fitness of each individual;
E, by { F (Sh), h=1,2 .., M } sorted according to ascending power, it is current preferably individual by the minimum individual mark of fitness value
Sbest, by the fitness value of minimum labeled as current preferably target function value Fbest;
F, according to the cumulative probability shown in formula (8) from P0Selection produces new colony P;
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>S</mi>
<mi>h</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>h</mi>
</munderover>
<mi>F</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>S</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>M</mi>
</munderover>
<mi>F</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>S</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>,</mo>
<mi>h</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<mi>M</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>8</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, p (Sh) represent individual ShCumulative probability;
G, according to crossover probability pcTwo individual S are randomly choosed out from PiAnd Si+1, intersect according to formula (9) and produce new two
Body SniAnd Sn(i+1), SniAnd Sn(i+1)Original S is replaced respectivelyiAnd Si+1, other individuals keep constant in P, by new population now
Labeled as Pc={ Sch, h=1,2 ..., M }, wherein SchRepresent PcIn h-th individual;
<mrow>
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>S</mi>
<mrow>
<mi>n</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>=</mo>
<msub>
<mi>&alpha;S</mi>
<mrow>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>+</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mi>&alpha;</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>S</mi>
<mi>i</mi>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>S</mi>
<mrow>
<mi>n</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mo>=</mo>
<msub>
<mi>&alpha;S</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mi>&alpha;</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>S</mi>
<mrow>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
</mrow>
</mtd>
</mtr>
</mtable>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>9</mn>
<mo>)</mo>
</mrow>
</mrow>
Caused random number between wherein α represents 0 to 1;
H, according to mutation probability pmFrom PcThe middle individual S of selectionch, according to formula (10) to SciRandom variation is carried out, produces new
Body SNh, and by SNhInstead of Sch, keep PcIn other individuals it is constant, so as to produce new population PN;
SNh=Sch+(r-0.5)×max{2(Sch-L),2(U-Sch)} (10)
Wherein L and U represent S respectivelychLower and upper limit, r represent 0 to 1 between caused random number;
I, P is unconditionally received0=PN;
J, repeat step D to step I is until meeting largest optimization iterations Imax;
K, the current preferably target function value F of outputbestWith best individual Sbest, and equivalent soil resistivity corresponding to output is pre-
Measured value ρ=yOl(Tmax);
L, 10kV overhead transmission line shaft tower grounding resistances R is calculated according to formula (11);
<mrow>
<mi>R</mi>
<mo>=</mo>
<mfrac>
<mi>&rho;</mi>
<mrow>
<mn>2</mn>
<mi>&pi;</mi>
<mi>L</mi>
</mrow>
</mfrac>
<mrow>
<mo>(</mo>
<mi>l</mi>
<mi>n</mi>
<mfrac>
<msup>
<mi>L</mi>
<mn>2</mn>
</msup>
<mrow>
<mi>h</mi>
<mi>d</mi>
</mrow>
</mfrac>
<mo>+</mo>
<msub>
<mi>A</mi>
<mi>t</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>11</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, ρ is the soil resistivity obtained in step (11), and L is 10kV overhead transmission line Reinforced Concrete Pole Tower earthing poles
Total length, h be earthing pole the depth of burying, d be earthing pole diameter, AtTo be grounded the form factor of level.
2. the mountain area 10kV line-to-ground resistive arrangement methods based on adaptive neural network as claimed in claim 1, it is special
Sign is:The step D assesses P in accordance with the following steps0In each individual fitness:
D1, make k=1;
D2, by tu、ρu、tm、ρmAnd ρlAs the input signal of adaptive neural network, by the input signal of kth time sampling instant
XI(k) it is defined as XI(k)={ xi(k), i=1,2 ..., nI, wherein nIFor neural network input layer nodes, in this nI=5,
x1(k)=tu,x2(k)=ρu,x3(k)=tm,x4(k)=ρm,x5(k)=ρl;
D3, the neutral net hidden layer input signal X for calculating kth time sampling instantHI(k)={ xHIj(k), j=1,2 ...,
nH, wherein xHIj(k) calculate as shown in formula (1):
<mrow>
<msub>
<mi>x</mi>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<msub>
<mi>n</mi>
<mi>I</mi>
</msub>
</munderover>
<msub>
<mi>w</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>,</mo>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>n</mi>
<mi>H</mi>
</msub>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, wij(k) connection weight of the kth time between sampling instant input node i and hidden layer node j is represented;D4, calculate the
The neutral net hidden layer output signal X of k sampling instantHO(k)={ xHOj(k), j=1,2 ..., nH, wherein xHjO(k) count
Calculate as shown in formula (2):
<mrow>
<msub>
<mi>x</mi>
<mrow>
<mi>H</mi>
<mi>O</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mn>1</mn>
<mo>+</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>H</mi>
<mi>I</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
</mrow>
</msup>
</mrow>
</mfrac>
<mo>,</mo>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>n</mi>
<mi>H</mi>
</msub>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
D5, the neutral net output layer signal Y for calculating kth time sampling instantO(k)={ yOl(k), l=1,2 ..., nO, wherein
nORepresent output layer number of nodes, yOl(k) calculate as shown in formula (3):
<mrow>
<msub>
<mi>y</mi>
<mrow>
<mi>O</mi>
<mi>l</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<msub>
<mi>n</mi>
<mi>H</mi>
</msub>
</munderover>
<msub>
<mi>w</mi>
<mrow>
<mi>j</mi>
<mi>l</mi>
</mrow>
</msub>
<msub>
<mi>x</mi>
<mrow>
<mi>H</mi>
<mi>O</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>,</mo>
<mi>l</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>n</mi>
<mi>O</mi>
</msub>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, wjl(k) connection weight between kth time sampling instant hidden layer node j and output node layer l is represented;
D6, assess according to formula (4) kth time sampling instant neutral net error performance index J (k);
Wherein, yOlr(k) desired value of the equivalent soil resistivity of kth time sampling instant is represented;
D7, the weight coefficient w according to formula (5) renewal+1 sampling instant of kthij(k+1);
<mrow>
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>w</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mi>&eta;</mi>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>l</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<msub>
<mi>n</mi>
<mi>O</mi>
</msub>
</munderover>
<mo>{</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>y</mi>
<mrow>
<mi>O</mi>
<mi>l</mi>
<mi>r</mi>
</mrow>
</msub>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
<mo>-</mo>
<msub>
<mi>y</mi>
<mrow>
<mi>O</mi>
<mi>l</mi>
</mrow>
</msub>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<msub>
<mi>w</mi>
<mrow>
<mi>j</mi>
<mi>l</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>x</mi>
<mrow>
<mi>H</mi>
<mi>O</mi>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>H</mi>
<mi>O</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>}</mo>
<mo>+</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>&beta;</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
<mo>-</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>(</mo>
<mrow>
<mi>k</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
<mo>)</mo>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
</mtable>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, η represents learning rate, and β represents factor of momentum;
D8, the weight coefficient w according to formula (6) renewal+1 sampling instant of kthjl(k+1), and k=k+1 is made;
wjl(k+1)=wjl(k)+η(yOlr(k)-yOl(k))xHOj(k)+β(wij(k)-wij(k-1)) (6)
D9, repeat step D2 are to step D8 until meeting k=Tmax, wherein TmaxRepresent maximum sampling number;
D10, according to formula (7) calculate individual ShFitness F (Sh);
<mrow>
<mi>F</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>S</mi>
<mi>h</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>k</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<msub>
<mi>T</mi>
<mi>max</mi>
</msub>
</munderover>
<mi>J</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>,</mo>
<mi>h</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<mi>M</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>7</mn>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710888308.8A CN107818199A (en) | 2017-09-27 | 2017-09-27 | Mountain area 10kV line-to-ground resistive arrangement methods based on adaptive neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710888308.8A CN107818199A (en) | 2017-09-27 | 2017-09-27 | Mountain area 10kV line-to-ground resistive arrangement methods based on adaptive neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107818199A true CN107818199A (en) | 2018-03-20 |
Family
ID=61607392
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710888308.8A Pending CN107818199A (en) | 2017-09-27 | 2017-09-27 | Mountain area 10kV line-to-ground resistive arrangement methods based on adaptive neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107818199A (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104834215A (en) * | 2015-03-24 | 2015-08-12 | 浙江师范大学 | Variation particle swarm optimized BP neural network proportion integration differentiation (PID) control algorithm |
CN105424760A (en) * | 2015-11-23 | 2016-03-23 | 西南林业大学 | Calibration method for soil resistivity and soil water content of rocky mountainous area |
CN106771616A (en) * | 2016-11-25 | 2017-05-31 | 烟台职业学院 | A kind of method for determining the equivalent soil resistivity of deep soil |
-
2017
- 2017-09-27 CN CN201710888308.8A patent/CN107818199A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104834215A (en) * | 2015-03-24 | 2015-08-12 | 浙江师范大学 | Variation particle swarm optimized BP neural network proportion integration differentiation (PID) control algorithm |
CN105424760A (en) * | 2015-11-23 | 2016-03-23 | 西南林业大学 | Calibration method for soil resistivity and soil water content of rocky mountainous area |
CN106771616A (en) * | 2016-11-25 | 2017-05-31 | 烟台职业学院 | A kind of method for determining the equivalent soil resistivity of deep soil |
Non-Patent Citations (3)
Title |
---|
方盼: "一类模糊神经网络的参数辨识和结构优化" * |
李增: "复杂土壤地区接地网电气特性的研究" * |
王攀等: "基于测井参数的煤系烃源岩总有机碳含量预测模型" * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102722909B (en) | Assembly line topology network dynamic simulation method based on adaptive-dimensional DEM (dynamic effect model) | |
CN103018604A (en) | Assessment method of electricity grid lightning strike risk and device thereof | |
CN113971463A (en) | Heat supply pipeline risk distribution analysis method and routing inspection path planning system | |
CN103163064B (en) | Fuzzy prediction method of corrosion rate of transformer substation grounding grid | |
CN109753703A (en) | A kind of distribution line lightning protection grade appraisal procedure | |
CN106295911B (en) | A kind of grid branch parameter evaluation method based on chromatographic assays | |
CN106250667A (en) | The monitoring method of a kind of landslide transition between states of paddling and device | |
Thapar et al. | Simplified equations for mesh and step voltages in an AC substation | |
CN108388707A (en) | D.C. magnetic biasing computational methods based on field circuit method under a kind of three-dimensional asymmetric structure soil model | |
CN115374682B (en) | Space-time collaborative high-precision curved surface modeling method and system | |
CN113591256B (en) | Mountain area transmission line tower lightning probability result evaluation method | |
CN109492756A (en) | More element conductor galloping method for early warning and relevant apparatus based on deep learning | |
CN108537376A (en) | Buried oil-gas pipeline soil corrosion prediction technique based on RS-PSO-GRNN | |
Gull et al. | Watershed models for assessment of hydrological behavior of the catchments: a comparative study | |
Nassery et al. | A proposed modelling towards the potential impacts of climate change on a semi-arid, small-scaled aquifer: a case study of Iran | |
Liu et al. | A new risk probability calculation method for urban ecological risk assessment | |
Kumar et al. | A macroscopic inception criterion for the upward leaders of natural lightning | |
CN111275292B (en) | Grounding grid state evaluation method based on fuzzy analytic hierarchy process | |
Hossain et al. | Design and optimization of substation grounding grid for ensuring the safety of personnel and equipment | |
CN107818199A (en) | Mountain area 10kV line-to-ground resistive arrangement methods based on adaptive neural network | |
Zhao et al. | Uncertainty assessment of mapping mercury contaminated soils of a rapidly industrializing city in the Yangtze River Delta of China using sequential indicator co-simulation | |
CN111597767A (en) | Random nanowire network topology analysis and electrical property simulation method | |
Androvitsaneas et al. | Artificial neural network methodology for the estimation of ground enhancing compounds resistance | |
CN110265906A (en) | A kind of grounding net of transformer substation state evaluating method and computer system | |
Dokur et al. | Finsler geometry for two-parameter Weibull distribution function |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |